Amiga OS 3.1 Source code leaked yesterday

[Minuous]

It is interesting to note that H&P apparently were only provided with V40 sources, that is very strange...

BTW I have found an incompatibility between intuition.library 42.86 and CygnusEd Pro 4.20.

[Olaf Barthel]

Originally Posted by Minuous

It is interesting to note that H&P apparently were only provided with V40 sources, that is very strange...

If I remember correctly, Haage & Partner only had very limited access to the v40 source code. The replacements for the preferences editors fell into that category, for example, which could not have been created without knowledge about the underlying framework. The bulk of the changes which were founded directly on official v40 source code came by Heinz Wrobel and myself. Workbench 3.5 had originally been intended to be an Amiga Technologies GmbH product, and by the time there was only Amiga International left standing which did not have an R&D section. Hence the deal with Haage & Partner.

BTW I have found an incompatibility between intuition.library 42.86 and CygnusEd Pro 4.20.

That and other surprises should be expected To the best of my knowledge work on this branch of intuition.library was never concluded, with the Commodore bancruptcy putting an end to development and testing. Also, in early 1994 modern RTG solutions (CyberGraphX, Picasso96) were still years away. At the time I believe there were only three contenders (Picasso II, Macrosystem Retina and EGS), and like its successors (CyberGraphX, Picasso96) these relied upon knowledge of register usage and data structure layout which does not necessarily still apply to the intuition.library 42.86 you tested.

During my own work, for example, I found that under my intuition.library port Picasso96 would allocate bitmaps only from chip memory, making the whole RTG system dog slow.

There are bound to be more surprises, not all of them fun.

[Olaf Barthel]

Originally Posted by RobertJDohnert

You can stand by it all you want but its wrong. Plain and simple.

Wrong how, wrong why?

You claimed that an admin within GitHub whom you are personally acquainted with placed the exact time of the DMCA notice and the takedown occured on Wednesday, January 6th.

I do not question that the takedown notice on Wednesday succeeded. The redacted takedown notice text suggests that it was Cloanto who eventually managed to disable the four repositories involved. I cannot verify that the takedown notice was served and the repositories were disabled on the same day.

My claim is that this was not the first DMCA takedown request to be submitted, that it was submitted by Hyperion on 31st December 2015 and that it was not processed until January 4th 2016 when it was rejected.

I have firsthand knowledge about this unsuccessful DMCA takedown notice. I put it to you that given that you have a source at GitHub, it could be possible for you to verify this claim I made. A DMCA takedown notice being what it is, the details and the e-mail correspondence between GitHub and Hyperion will be on file.

Secondly, when did you become a spokesman or an employee for Hyperion or Cloanto? I dont know what kind of strange perversion you are getting out of this whole ordeal but its over. Really this whole leak is none of your business or non of your affair. If you turned in Github or any of the other 10 download sites you can get it from now, good for you. You deserve a lollipop. But otherwise man get on with real life. I sure as hell wouldnt be making defensive statements for either of those companies unless they paid me a crap ton of money to do it.

Ick. This is unnecessarily angry and rude.

It seems your account on this forum was created very recently, you only posted three times, and the wording of your message attempts to stir up trouble. Sir, I believe you are trolling me.

For what it's worth:

You are asking the wrong question. I have worked as a consultant in this crazy Amiga business since 1989, never as an employee. I even made money in this business, never burnt bridges, tried hard to be good guy (which sounds easier than it really is, believe me). As far as I know Hyperion and Cloanto do not have spokesmen, the respective busines owners handle these affairs all on their own. These are small operations after all.

I am the uncommon type of Amiga person who does have a dog in this fight, so to speak, and who has responsibilities to other persons involved in the same line of work. This is why I chose to comment on the matter at hand, which I believe is appropriate.

I was not involved in going after the download sites whose number you place at 10 (seems awfully low to me). Considering how much work must have been involved in that task, and me not even lending a hand, I would say that I would not deserve even a lollipop.

I am getting on with real life, thank you for mentioning it. Could be better, could be worse, I suppose. I (uncharacteristically) cheerfully choose to believe that it really could be worse, and hope that it will get better still.

It is possible that Amiga users exist who do not feel contempt or pity for Hyperion or Cloanto, and holding this point of view does not necessarily require money to be paid. This is a small market and margins are thin, enough said. Whether or not you want to be part of such a group, or rather the other group, or rather the group who does not care either way is, of course, your prerogative. As is the amount of money you could hope to ask either Hyperion or Cloanto to pay you in order to influence your decision.

[ptyerman]

Well said Olaf. +1

[tolkien]

Olaf +2 good words.

[kolla]

All this drama is a play for the gallery, so that everyone can tap their own shoulders and say "we did what we could do" and claim contract agreements are upheld. Reality is of course that the sources are for all foreseeable future available for anyone who may find interest in them. And there are plenty of people who do, and who use this knowledge to improve their products, as they should do.

[A1200]

Regarding OS3.5/3.9 I liked them and the bundled software but they made an Amiga with modest accelerator a lot slower, so I run my A1200/28/6MB RAM with 3.1 and patches.

[copse]

Originally Posted by kolla

All this drama is a play for the gallery, so that everyone can tap their own shoulders and say "we did what we could do" and claim contract agreements are upheld. Reality is of course that the sources are for all foreseeable future available for anyone who may find interest in them. And there are plenty of people who do, and who use this knowledge to improve their products, as they should do.

It'd be interesting to see a list of all leaked retro-computing related source codes. I had been hoping we'd see a flow from the psygnosis disks we heard about several months ago.

[masteripper]

Sorry to dig up such an old thread but I am curious ... consider it's size (doing the sum) is it the real thing?
If I am not mistaken kickstart is 512k and WB 6 disks x 880k ...so how it's such big ... several megabytes..
Has actually anyone has some info if it was compiled and produced the "real" thing....

[redblade]

It was massive. I think over 100mb unpacked?? (For an amiga is huge!!)

Came with all the source to kick 3.1. I can't remember seeing the source to 1.3 in there.
I remember seeing the source to Tools/Commodities/screen blanker (I hope they could release that, or part of it as a template source, so coding screen savers would be a bit easier for copy and paste coders).

I also remember seeing some Japanese (Hiragana) stuff as they were working on making 3.1 international.

That's all I remember. Was years ago.

[PeterK]

It's a 7z with 63 MB and contains nearly 500 MB and over 46000 files, but I've never tried to compile anything from that source. It costs too much time to examine and understand the code and I couldn't find anything useful for me. Maybe something for the OS developers who will have that source anyway.

[gulliver]

For us there is not much value in there:

The build scripts are broken, some components missing, also some of the compilers are missing.

In addition to that, the code present there is nowadays extremely outdated as we have fixed a huge amount of bugs and made lots of enhacements when we realeased 3.1.4 then its update 3.1.4.1 and are now polishing 3.2.

However, it is good from a historical (archeological?) point of view, if you enjoy that type of journey. :-)

You get to understand where the OS modularity comes from, and the surprsingly difficult jumps and hops developers had to overcome to get something out to users.

[masteripper]

Originally Posted by gulliver

For us there is not much value in there:

The build scripts are broken, some components missing, also some of the compilers are missing.

In addition to that, the code present there is nowadays extremely outdated as we have fixed a huge amount of bugs and made lots of enhacements when we realeased 3.1.4 then its update 3.1.4.1 and are now polishing 3.2.

However, it is good from a historical (archeological?) point of view, if you enjoy that type of journey. :-)

You get to understand where the OS modularity comes from, and the surprsingly difficult jumps and hops developers had to overcome to get something out to users.

Thanks for the info...
I read a lot of posts online about how things could have being different if this original leak (along with the previous Amithlon) was allowed to circulate...
If i read correctly given the breakthrough design of Amiga/Os it would easily gain a place in the current scene of OS along with Windows/Linux and the derivatives (Android/MacOs....etc)..it just needs to reach the x86/Arm platforms and get away from the shrinking 68k/PPC ecosystem.

[absence]

Originally Posted by masteripper

If i read correctly given the breakthrough design of Amiga/Os it would easily gain a place in the current scene of OS

That's a bit of a misunderstanding I'm afraid. Amiga OS is a hodgepodge of shoestring, bubblegum, and speed hacks, thrown together in a hurry to make it to the market in time. Don't get me wrong, they did very well for a mid-80s consumer operating system given their constraints, but the design is ill suited to evolve into a modern operating system. Doing so would break all kinds of compatibility, essentially making it a new operating system in all but name. There are already several such alternative operating systems available, without mainstream success.

Originally Posted by masteripper

it just needs to reach the x86/Arm platforms and get away from the shrinking 68k/PPC ecosystem

Are you aware of AROS and the Icaros distribution? It's an open source reimplementation of Amiga OS that seems quite capable from the screenshots.

[Olaf Barthel]

Originally Posted by absence

That's a bit of a misunderstanding I'm afraid. Amiga OS is a hodgepodge of shoestring, bubblegum, and speed hacks, thrown together in a hurry to make it to the market in time.

You sound like you're romanticizing the result

Seriously, the original Amiga operating system design is a mixed bag indeed. On the whole the good outshines the poorer and the wobblier parts, in my opinion at least. While the decisions which led to the design of individual components and their programming interfaces have not all aged well, the foundations, especially in context, have stood the test of time.

By this I mean the extendability built into exec.library and everything which sits on top of it. This was an open design which time and again allowed the operating system to be adapted and enhanced, even if you did not have access to the source code. This sort of thing was rare at the time it was designed. Today's mainstream operating systems share similar aspects, although the extendability is now considered at least a security risk, if not worse...

That said, the API design isn't always so great, on average. graphics.library in particular has not aged so gracefully, for example. It was not designed to be extended, and that goes right down to the data structures it uses. Documentation has always been spotty and poor, but, hopefully, that may be addressed in the NDK 3.2 work (or we'll at least make a valiant attempt to get it started).

Speaking of graphics.library, I think there's something to be said in favour of how much effort was spent on making the custom chips' power accessible through its APIs. It's no mean feat to design something like BltBitMap() for the blitter hardware and have that very same architectural pattern show up in the entire API stack that built upon the graphics primitives.

In hindsight, it might have helped if parameter validation and error reporting had been better throughout the operating system. Interfaces to hardware features were always much closer to the hardware than to providing an abstraction which could be retooled and adapted in the future. For example, the audio.device is a very thin layer on top of the audio hardware, while what could have been more impactful would have allowed you to play back sound with a dedicated API that could have employed mixing and playback rate conversion.

What we got was so much better than the alternatives on offer at the same time. Compare this to how the Apple Macintosh fared in the same time frame. That was a very tightly closed system and its operating system design was very in much alignment with the best practices for this kind of machine. Not so the Amiga...

[absence]

Originally Posted by Olaf Barthel

You sound like you're romanticizing the result

It was an enjoyable hodgepodge after all!

[masteripper]

Originally Posted by absence

That's a bit of a misunderstanding I'm afraid. Amiga OS is a hodgepodge of shoestring, bubblegum, and speed hacks, thrown together in a hurry to make it to the market in time.

Although so many mistakes where made still the year is 2020 ...the original machine is 35 years old and without proper inside knowledge we have managed to have Full HD desktops with a ton of applications waiting to get the extra power . (maybe i am cheating a bit since many of extras came from later "extra" but the year is somewhere in '90s)
I don't think that other OSes could have such an extensibility and given the fact that somewhat the whole scene is "abandoned" it would be something for future history books "a what if.....".
As an original Amiga user back then...i reckon that if the dread Guru meditation error was handled as it should (i read that some guy was about to fix this but the bosses of Commodore just refused to pay him the money and so.....poor Amiga.) and the Imaging department was more flexible (stupid interlace)...things would be far different....

[absence]

Originally Posted by masteripper

Full HD desktops with a ton of applications waiting to get the extra power . (maybe i am cheating a bit since many of extras came from later "extra" but the year is somewhere in '90s)
I don't think that other OSes could have such an extensibility

It's funny that you should mention RTG, because competing operating systems from the same era as Amiga OS 3.0, like Windows 3.1 and Apple's System 7.1, already supported this natively without hacks like P96. They had their own flaws for sure, but were really quite capable, both in native features and extensibility.

[Thomas Richter]

Originally Posted by masteripper

I don't think that other OSes could have such an extensibility and given the fact that somewhat the whole scene is "abandoned" it would be something for future history books "a what if.....".

I am not quite sure what you are trying to imply here. Other systems are equally extensive. Solaris had ".so"s instead of shared libraries, windows has dlls. Other Oses made less mistakes in the low level design especially around resource management and graphics. Some made more mistakes in the multitasking domain. AmigaOs is a consumer Os, and not a robust design.

Originally Posted by masteripper

As an original Amiga user back then...i reckon that if the dread Guru meditation error was handled as it should (i read that some guy was about to fix this but the bosses of Commodore just refused to pay him the money and so.....poor Amiga.) and the Imaging department was more flexible (stupid interlace)...things would be far different....

Well, but that means throwing the current design away and re-do it. The problems you mention are very fundamental design errors, and not just minor inconvenience issues.

[Thomas Richter]

Originally Posted by absence

It's funny that you should mention RTG, because competing operating systems from the same era as Amiga OS 3.0, like Windows 3.1 and Apple's System 7.1, already supported this natively without hacks like P96. They had their own flaws for sure, but were really quite capable, both in native features and extensibility.

The trouble with the Amiga "design" is that graphics is a rather thin layer around the custom hardware, and its capabilities, that is, lack of proper abstraction, and that it lacks proper interfaces, making most of its structures public.


You even see cases where other system components replicated structures of graphics. The whole region/rectangle logic of graphics is replicated within layers, just as ClipRects. The monitor layer of graphics conflicts and partially replicates the display info database. Actually, the latter is just a quite awkward design, and there are some traces in the Os that this was going to be replaced by something better.


Also, CBM made some attempts in 3.1 to at least allow third party RTG systems, though the work is incomplete and it still (even in 3.2) needs some hacks to get RTG working. I am currently going through a review of P96 just to know what is needed to make this a "working system", and this goes unfortunately beyond graphics in some corner cases.